[This application has been revised and substantively improved on the basis of reviewers' recommendations. Changes within the proposal are highlighted by a different font (compared to this original font) and are bracketed. The following improvements should be apparent in the present proposal: Many studies from the original proposal have been dropped, and the project has been streamHned. It is difficult to resist being overly ambitious, as was true of the original submission, when proposing five years of experiments from a large and productive research team that is excited about our collaborative opportunities, both among ourselves and with the other investigators in this program who are studying complementary phenomena. On the one hand, the rhesus monkeys tested in these studies complete, as a group, almost 100,000 trials per week across tasks and studies, providing an ample foundation for the present investigations as well as the studies proposed by other researchers in this program. Similarly, this project team tested an average of over 500 undergraduate volunteers per year in the last four years. We certainly want to generate the most science possible across the proposed 5-year funding period. On the other hand, we acknowledge that the more studies that are described in a single proposal, the less clear the details of those studies can be, the less coherent the proposal appears, and the harder it becomes to see the theme that ties the studies together. In organizing the remaining experiments, we have sharpened our construct definitions by ensuring that the tasks in Study 1 reflect the control of attention (selection of some stimuli rather than others) for processing. There is of course a longstanding debate regarding whether attention is selection of stimuli for processing (early selection), or selection of a response (late selection). In light of the reviewers' comments, we avoided this debate in the present proposal by choosing early-selection tasks (tests of how well individuals select some stimuli and ignore others). We moved response-selection tasks, together with other tests of cognitive control, into Study 2. Several very interesting cognitive tests related to executive functioning (including planning, monitoring, and statistical learning were deleted from Studies 2 and 3 so as to maintain theme of the control of attention across the project. Although we agree with the reviewer who described these tasks as clever and compelling, we also agreed with the reviewers who saw them as peripheral to the central theme. Study 3 was consequently refocused on the reciprocal role of attention and learningwhich seems critical, given our desire to understand how learning establishes the experiential and executive constraints that vie for control over attention (i.e., of anchoring executive in behavior rather than allowing it to remain an undefined homunculus. This study also supports our translational effort to identify particular t3rpes of training (including symbol training) that might alter the control of attention. The net result of these reductions in experiments, together with the decision to move the details of the fMRI testing and analyses to the Core where they belong, provided space for elaborating on the brain-behavior experiments we will tackle during this funding period. We have attempted to show that it is timely to study these cognitive competencies using neuroimaging technologies. We have also added to the preliminary studies to build the foundation for this entire proposal. At the same time, we did add several experiments that were specifically recommended by the reviewers or by our review of the recent literature, including (a) a replication of two previous findings (Ei.i) using conditions that calibrate baseline performance levels across species; (b) addition of the eyes-looking task to complete the possible comparisons within the cognitive-control study E2.1; and (c) addition of CPT-AX testing to distinguish between proactive and reactive responding, consistent with a recent model of cognitive control. The results of Ei.i could potentially change the design of each subsequent experiment. In this revision, we believe that we've achieved a good balance by including the studies that have the greatest probability of addressing our specific aims, and by eliminating relatively uninformative or potentially redundant tests of specific populations or with specific measures. We recognize that one confusing element of the original proposal pertained to which participant groups would be tested on which specific studies. In part, this is a result of our desire to base those decisions on the results of earlier experiments. For example, we don't want to administer a task to children or chimpanzees until it has been shown to be a good task, in the sense of producing meaningful variations as a result of the independent-variable manipulations, in testing with undergraduate participants or monkeys. It is neither practical nor scientifically necessary to test every group (naive and experienced macaques, capuchins, chimpanzees, children, undergraduates with and without ADHD) on every task and condition. However, we want understand attention control from comparative, developmental, neuropsychological, and of course cognitive perspectives, and thus it is necessary to test multiple groups. Additionally, to control for the influence of different levels of motivation, training, and so forth, it is necessary to produce converging evidence by using multiple tasks. Taking all of this into consideration, the final proposal is summarized below: Study 1. The determinants of attention: What controls selection of cues that compete for processing? Subjects = Rhesus monkeys. Capuchin monkeys, undergraduate volunteers; chimpanzees, children and adults with ADHD possible on a subset of tasks, contingent on initial findings; neuroimaging studies likely with a subset of tasks, contingent on initial findings (e.g., fMRI and TCD with humans of Stroop-like selection, ANT, CPT; TMS witii monkeys of anti-saccade, ANT) a. Stroop-like selection: Ei.i=numerical Stroop; Ei.3=ANT; Ei.4= multi-modal Stroop, social Stroop (E1.4), global/local, bullseye flanker b. Attention scanning: Ei.i=anti-saccade; Ei.2=dual-task paradigms; Ei.3=ANT c. Attention sustaining: Ei.3=CPT, MOT, ANT Study 2. How does attention control relate to other aspects of cognitive control? Subjects = Rhesus monkeys. Capuchin monkeys, undergraduate volunteers; chimpanzees, children and adults with ADHD possible on a subset of tasks, contingent on initial findings;; neuroimaging studies possible with a subset of tasks, contingent on initial findings (e.g., fMRI and TCD with humans of inhibition tasks, running memory; TMS with monkeys of Simon-task) a. Inhibition tasks: E2.i=dots, heart/flower, eyes-looking; E2.4=stop-signal b. Set-switching: E2.2=Shape School, WCST c. Working memory: E2.3=running memory, N-back, symmetry span Study 3. How does learning influence attention control, and how does attention influence learning? Subjects=Rhesus monkeys, Capuchin monkeys; humans and chimpanzees possible , contingent on initial findings a. Relational/Associative learning: E3.i=meaningful failures, meditational paradigm b. Training effects: E3.2=symbol training; E3.3= executive training The theme that binds these studies is the competition between stimulus events (e.g., attention capture), stimulus associations (e.g., conditioned or primed attention), and higher-order intentions (e.g., executive attention) for the control of selection and behavior. Although this specific model is not accepted in the literature, there is little reason for concern that the data would be impugned if the framework is rejected. This theoretical perspective echoes the longstanding and widely embraced distinction between top-down and bottom-up processing (a distinction known by many names) and is consistent with the recent effort in comparative cognition to understand cognitive control while acknowledging stimulus control. Our competition framework, inspired by neural-net connectionist modeling, finds theoretical kinship with numerous other theories (e.g., the race model of attention; Bundesen, 2000). Although the distinction between the capture of attention by environmental cues (like motion, sudden change, and dishabituation) and control of attention by experience (as in contention-scheduled, automatic, and primed processing) is unique to the present model, the experiments proposed here will permit empirical test of whether attention is influenced by variables falling into these three specific classesor perhaps into just two, or even just one category. In summary, one need not embrace the framework to find value in the studies. What seems unlikely to be fragile about the model is the assumption that multiple potential cues compete for attention (as reflected behavior) at any moment in time; that by varjdng the potency or strength of each of these cues and measuring the results on response latency, accuracy, and pattern, one can identify meaningful individual and group (including species) differences in the control of attention; that these characteristic differences in sensitivity to various competing constraints on attention may also be evident in the control of other cognitive operations; and that reliable variations in the control of attention should be manifest in different patterns of brain activity that correspond to the different patterns of behavioral response. At a large and diverse university like Georgia State where most of the human participants will be tested, concerns about the representativeness of the sample are limited. However, this proposal also extensively employs a colony of highly experienced resident animals at the Language Research Center, and so we are particularly concerned about the suggestion that these animals and their prior experience could seriously compromise the present results. Without doubt, we are able to test these monkeys and apes on tasks that would be difficult or impossible to administer to naive animals. Similarly, these animals have demonstrated some cognitive competencies that were heretofore thought beyond the range for monkeys or apes. Part of the reason we have such confidence in Our hypotheses that monkeys differ from humans in the capacity for executive control of attention is that these particular monkeys seem ideally trained and prepared for comparison with humans on computerized cognitive tasks. Each of the tasks proposed here builds on extant repertoires and prior experience, just as we assume that the humans will enter each test with a history of experiences in attending, learning, and problem solving that serves to prepare them. That said, we do intend to assess the role of experience, in part by testing relatively naive macaques on selected tasks. In sum, we believe that the monkeys' and chimpanzees' history of participation in cognitive research supports and is a strength of the current proposal. As the reviewers correctly note, the critical challenge of such a program of researchindeed, of every scientific study of different groups, whether the groups are defined by species, age, culture, diagnostic category, performance level, or some other criterionis to ensure that the different groups are tested comparably. This team of investigators is highly experienced with such between-groups comparisons, although this alone does not change the difficulty of the task at hand. Individual and group differences in sensitivity to stimulus conditions, delay of reward, and similar variables are inherent in these comparisons, and indeed are a topic of study in this proposal. We have built several validity- and calibration-checks into these studies, and we've expanded the discussion of these in the proposal. Briefly, our confidence in the conclusions from these between-groups comparisons will be increased to the degree that (a) they reflect convergent evidence across tasks and manipulations (e.g., species differences in the capacity for the executive control of attention are seen in Stroop and vigilance tasks, and are similar for manipulations of incentive and for manipulations of concurrent workload); (b) within-subject differences serve as the foundation for between-groups comparisons (e.g., monkeys are not less attentive than chimpanzees, but compared to chimps the monkeys' attention was more affected by increases in stimulus-response association strength); (c) manipulations have similar effects across groups in baseline performanceverifying that performance is not at ceiling or floor and that the manipulation is sufficiently large to influence behaviorbut different effects across groups in the critical stimulus-conflict conditions (e.g., stimulus movement improves target detection for children and adults, but is more disruptive for children than adults when nontarget stimuli move); and (d) training and/or instructions are provided to ensure that asymptotic or criterial performance is being compared for all groups. Without denjdng the challenge before us, we believe that the reviewers' suggestions have improved our ability to address our specific aims with studies that will withstand the judgment of the literature.